Administration of all-transretinoic acid (ATRA) + myelotoxic chemotherapy results in long-term remission in 70% of patients with newly diagnosed acute promyelocytic leukemia (APL). It is becoming clear however that this approach is associated with development, several years later, of myelodysplastic syndromes and AML. The demonstration of the effectiveness of arsenic trioxide (ATO) in APL makes it feasible to assess, in newly diagnosed APL, whether the combination of ATO + ATRA will enable elimination of myelotoxic therapy. To test this hypothesis (SA#1) we will conduct a trial of ATO + ATRA, with myelotoxic therapy added only if minimal residual disease (MRD), as judged by the standard manual PCR assay, persists or recurs. For safety monitoring we will use a published Bayesian "multiple outcome" design that allows early termination if the rates of either CR, or PCR negativity at 6 months from CR date, are too low. More effective means of measuring MRD would obviously make similar trials more feasible in the future, and SA#2 tests the hypothesis that use of high sensitivity quantitative real-time PCR, rather than the standard assay, and blood rather than marrow will increase the accuracy of current methods. Similarly, understanding of mechanisms underlying resistance to ATRA would increase the possibility of eliminating myelotoxic therapy, and SA#3 tests the hypothesis that addition of ATO to ATRA, while decreasing the overall relapse rate, increases the frequency of missense mutations in the PML-RAR( gene among patients who do relapse. [unreadable] [unreadable]